Aiming at the defects of late convergence speed, low search accuracy and low stability of wild horse optimizer, a hybrid multi-strategy improved wild horse optimizer was proposed. Firstly, Halton sequence initialization was used to increase population diversity. Secondly, the adaptive parameter was improved to balance the global search and local development capabilities. Then, the worst position of individual population was improved by simplex method. Finally, the escaping behavior was added to improve the optimization accuracy of the algorithm. In order to verify the effectiveness of the improved strategy, 9 standard test functions are selected for simulation experiments. The improved algorithm is applied to mechanical design problems and truss structure optimization examples, and the optimization results are reduced by 16.61%, 0.21%, 2.96% and 0.61% compared with the original algorithm. The statistical results show that the improved algorithm has higher optimization accuracy than the basic algorithm and other comparisons in solving practical engineering problems.